Re: [swanmodel-users] Trouble with unstructured mesh

[Amin I. <ami...@gm...>]

Dear Paulo Avilez-Valente,

It seems you are trying to run unstructured SWAN in parallel mode. You need
to use "fort.14" grid (ADCIRC) format to be able to run unstructured SWAN
in parallel mode. You also need to compile SWAN+ADCIRC and use adcprep to
split your grid.  However, you should be able to run SWAN using TRIANGLE
grid in single mode.

Regard,
Amin

On Mon, Mar 11, 2019 at 12:52 PM Paulo Avilez-Valente <pva...@fe...>
wrote:

> Dear all
>
> I have a case to analyze where I need to use an unsctructured mesh (SWAN
> version 41.20AB)
> I used TRIANGLE to produce the mesh. Afterwards I edited the my_mesh.node
> and my_mesh.ele files to format the coordinates, so tabs were introduced.
> This was done in Windows 7. SWAN was run in a Linux machine.
>
> I used the SWAN commands:
> CGRID UNSTRUC CIRCLE 72 .04 .5
> READGRID UNSTRUC TRIANGLE 'my_mesh'
>
> when running SWAN and after the command
> BOUND SHAPESPEC JONSWAP 3.3 PEAK DSPR DEGREES
> SWAN stops with the following message
>
> ** Terminating error: SwanBpntlist: list of boundary vertices could not be
> completed
>
> I checked the "swanmodel-users list" and it seems that the last time this
> error was referred to was in 2012, when it seemed to result from a
> non-optimal algorithm for boundary identification. From the discussion in
> 2012, I thought that the problem had been solved in the code.
>
> My mesh seemed to be fine, but I noticed that the sequence of nodes in
> some triangles were not counter-clockwise (ccw) oriented. This happened to
> about 40% of the total number of triangles. I edited the my_mesh.ele file
> so that all triangles are now ccw oriented.
>
> Unfortunately this did not improve the outcome.
>
> Now SWAN stops immediatly after the command
> CGRID UNSTRUC CIRCLE 72 .04 .5
> It writes to the PRINT file:
> MDC=    72  (the rest of the information on the frequency doamin
> discretization is missing)
> and stops with the following message to the monitor:
>
> Program received signal SIGSEGV: Segmentation fault - invalid memory
> reference.
>
> Backtrace for this error:
> #0  0x7fd2ae76931a
> #1  0x7fd2ae768503
> #2  0x7fd2adde5f1f
> #3  0x562841b9054e
> #4  0x562841b8e74d
> #5  0x562841ae9648
> #6  0x562841ad74c9
> #7  0x562841ad9220
> #8  0x7fd2addc8b96
> #9  0x562841a8ad89
> #10  0xffffffffffffffff
> Segmentation fault (core dumped)
>
> So, do you have any idea aboutwhat is causing this?
>
> The user manual is not very keen on information about the format that is
> expected for the TRIANGLE files.
>
> TRIANGLE produces 2D meshes. Nevertheless, if an attribute is used in the
> *.node file (beside the boundary marker), a 3D surface can be generated.
>
> For a mesh with 1000 nodes, which of these two "my_mesh.node" files shall
> be used? "my_mesh_01.node" ou "my_mesh_02.node"? Or are both acceptable?
>
> my_mesh_01.node
>
> 1000 2 0 1
> 1 1000.00 500.00 2
> 2 1105.00 530.00 2
> 3 1213.00 555.00 1
> ...
> 1000 2035.00 700.00 0
>
> my_mesh_02.node
>
> 1000 2 1 1
> 1 1000.00 500.00 102.45 2
> 2 1105.00 530.00 155.34 2
> 3 1213.00 555.00 311.05 1
> ...
> 1000 2035.00 700.00 225.34 0
>
>
> The same goes for the mesh topology file "my_mesh.ele". Can we use a file
> that has an attribute for each element? Or shall we not include attributes?
>
> For a mesh with 1500 triangles, which of these two "my_mesh.ele" files
> shall be used? "my_mesh_01.ele" ou "my_mesh_02.ele"? Or are both acceptable?
>
> my_mesh_01.ele
>
> 1500  3 0
> 1 5 7 10
> 2 11 9 8
> 3 100 245 10
> ...
> 1500 13245 2345 9875
>
> my_mesh_02.ele
>
> 1500  3 1
> 1 5 7 10 1000.00
> 2 11 9 8 1200.00
> 3 100 245 10 1000.00
> ...
> 1500 13245 2345 9875 1000.00
>
>
> Finally the file with the depth information.
>
> On my first trials I used a regular input bottom grid with a flat bottom.
> But I wish to provide the depth information for each node.
>
> I intend to use the command
>
> INPGRID BOTTOM UNSTRUCT
> READINP BOTTOM 1. 'my_mesh.bot' FREE
>
> Is it correct? Shall I include the FREE keyword?
> But what is the format that shall be used in the "my_mesh.bot" file? Of
> the two following examples, "my_mesh_01.bot" and my_mesh_02.bot" which one
> is the correct? Or are they both wrong?
>
> my_mesh_01.bot
>
> 102.45
> 155.34
> 311.05
> ...
> 225.34
>
> my_mesh_02.bot
>
> 1000
> 1 102.45
> 2 155.34
> 3 311.05
> ...
> 1000 225.34
>
> I hope someone can give me a clue on any of these subjects.
>
> Best regards
>
> Paulo Avilez-Valente
>
>
> Dr.-Eng. Paulo Avilez-Valente
> Assistant Professor
> Faculdade de Engenharia
> Universidade do Porto
> Rua Dr. Roberto Frias
> P-4200-465 Porto
> Portugal
> _______________________________________________
> swanmodel-users mailing list
> swa...@li...
> https://lists.sourceforge.net/lists/listinfo/swanmodel-users
>


-- 
*Amin Ilia*